laravel-ai-sdk Eval report.

Validation plan

What the agent should do next

1. 1Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. 2Install in an isolated workspace or sandbox with no production secrets available.
3. 3Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. 4Compare the selected skill against at least one alternative when the eval status is review or failed.
5. 5Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Do not use when

Conditions that require another skill

• teams that need a vendor-supported SLA
• production agents without a repository review
• SKILL.md lacks a dedicated setup/prerequisites section covering installation and provider configuration.
• No explicit limitations or safety boundaries are documented for AI tool usage, such as restricting dangerous tools or validating external content.
• The skill is largely a documentation reference; it may require additional context for agents to know when to apply specific rules in a real workflow.
• Quality score needs review